Mira Murati launches Thinking Machines Lab with stacked OpenAI alumni team

Summary

Mira Murati launches Thinking Machines with OpenAI alumni team, over $100M in funding

Mira Murati, former CTO of OpenAI, has launched Thinking Machines Lab, an AI research and product company backed by over $100 million in venture funding. The company's website emphasizes "AI that works for everyone" and human-AI collaboration, positioning itself around three core gaps: the scientific community lags behind frontier AI capabilities, knowledge of how top systems are trained remains concentrated in a few labs, and existing AI systems resist customization to individual needs and values.

The team is exceptionally stacked with OpenAI alumni. Alex Gartrell led server operating systems at Meta with Linux expertise. Alexander co-created advanced voice mode at OpenAI and multimodal models at Meta. Andrew Tullock worked on ML systems at both OpenAI and Meta. Barrett Zoff, the CTO, is a former VP of research and post-training at OpenAI and co-creator of ChatGPT. Bryden worked on post-training research at OpenAI. Luke Metz, listed deep in the team roster despite being a co-creator of ChatGPT, represents the depth of talent on board.

The company's public positioning is notably bland—the website copy reads interchangeably with any other foundation model startup, describing flexible, adaptable, and personalized systems without articulating what differentiates Thinking Machines. The team did not bother setting up proper domain routing (thinkingmachines.ai without www), suggesting either deliberate indifference to marketing or the conviction that the brand itself carries sufficient weight.

Questions about differentiation linger. Whether the engineering pedigree translates to advantage remains unclear. The model company category is already crowded—Ilya Sutskever is raising $1 billion with $500 million from Greyscale Partners; xAI, Anthropic, and others are all staffed with top-tier researchers claiming they can ship frontier models in compressed timelines. The experience of Facebook's dominance despite Google's multi-year effort to dislodge it with Google+ suggests that team quality, while necessary, does not guarantee winner-take-most outcomes. The revealed preference of Murati and other OpenAI alumni to start separate companies rather than consolidate under a single "anti-Sam Altman" organization suggests the motivation is equity and control over shared mission.

On capital and compute strategy, the details are sparse. Murati likely has relationships with hyperscalers willing to provide GPU access and training infrastructure. Amazon, which maintains loose alignment with Anthropic but no exclusive relationship, could benefit from additional leverage by backing Thinking Machines as a hedge against OpenAI dominance. How Thinking Machines sources compute—whether it builds its own cluster, partners with a cloud provider, or pursues a hybrid model—remains unstated.

The company says it will publish technical papers and code frequently but does not commit to open-sourcing its models, which undermines the stated vision of democratizing AI knowledge. That distinction suggests the open-science framing is tactical positioning rather than operational commitment.

On Murati's background: She was born in Albania, moved to Canada as a teenager, studied mechanical engineering at Dartmouth, and led product and engineering at Leap Motion, the VR hand-tracking startup whose founder David Holz went on to create Midjourney. She spent time at Tesla as a senior product manager on the Model X. She joined OpenAI in 2018 as VP of applied AI and partnerships, was promoted to senior VP in 2020, and became CTO in 2022. She briefly served as interim CEO during the November 2023 board upheaval that ousted Sam Altman. The Guardian profile notes she has consistently framed AI as a tool, echoing Altman's public positioning, and told Fortune she was drawn to the possibility of achieving general intelligence.

On Microsoft's absence: Satya Nadella publicly praised Murati's work building "some of the most exciting AI technologies we've ever seen," but there is no indication Microsoft invested in Thinking Machines. At the time of her departure, Nadella appeared more aligned with OpenAI. More recently, Microsoft has shown willingness to make direct investments in potential competitors, given that OpenAI now works with Oracle and others competitive to Microsoft.

The central question is not whether Murati can build a frontier model—her track record and team suggest she can within roughly a year—but whether the foundation model category itself is commoditizing, and whether a third or fourth well-funded entrant can capture meaningful economic value in a power-law winner-take-most market.